Speech-recognition circuitry employing nonlinear processing, speech element modeling and phoneme estimation

Abstract

A phoneme estimator in a speech-recognition system includes energy detect circuitry for detecting the segments of a speech signal that should be analyzed for phoneme content. Speech-element processors then process the speech signal segments, calculating nonlinear (powers and products) representations of the segments. The nonlinear representation data is applied to speech-element modeling circuitry which reduces the data through speech element specific modeling. The reduced data are then subjected to further nonlinear processing. The results of the further nonlinear processing are again applied to speech-element modeling circuitry, producing phoneme isotype estimates. The phoneme isotype estimates are rearranged and consolidated, that is, the estimates are uniformly labeled and duplicate estimates are consolidated, forming estimates of words or phrases containing minimal numbers of phonemes. The estimates may then be compared with stored words or phrases to determine what was spoken.

What is claimed as new and desired to be secured by Letters Patent of the United States is: 
     
       1. A speech-recognition device for identifying a speech-element of interest in a speech signal, said device comprising: A. processing means for processing the speech signal to produce for signal segments a reduced-data representation of the speech that includes a plurality of reduced-data elements;   B. a first means for calculating quantities proportional to products of certain of the reduced-data elements and power of certain of the reduced-data elements to produce a nonlinear representation of the speech that includes as elements thereof the quantities proportional to the products and powers;   C. first modeling means for modeling the nonlinear representation with a group of modeling elements which include nonlinear representations characteristic of one or more speech-elements of interest in known speech, and producing a reduced-data nonlinear representation that includes a plurality of reduced-data nonlinear representation data elements;   D. a second means for calculating quantities proportional to products of certain of the reduced nonlinear representation data elements and powers of certain of the reduced nonlinear representation data elements to produce a further nonlinear representation of the speech that includes as elements thereof the quantities proportional to the products and powers;   E. second modeling means for modeling the further nonlinear representation with a group of modeling elements which include nonlinear representations characteristic of the speech-elements of interest in known speech, the modeling producing data which identifies the speech elements present in the speech signal.   
     
     
       2. The speech-recognition device of claim 1, wherein said second calculating means includes means for concatenating the reduced nonlinear representation data elements corresponding to a predetermined number of signal segments before calculating the proportional quantities. 
     
     
       3. The speech-recognition device of claim 1, wherein the second nonlinear calculating means includes a nonlinear receptive processor which forms products of selected data associated with the energy of the speech signal in various frequencies and selectively combines the products to produce data which relate, respectively, to changes in the signal energy signal over various frequencies and over predetermined time periods. 
     
     
       4. A speech-recognition device for identifying a speech-element of interest in a speech signal, said device comprising: A. processing means for processing the speech signal to produce a reduced-data representation of the speech that includes a plurality of reduced-data elements;   B. a first means for calculating quantities proportional to products of certain of the reduced-data elements and powers of certain of the reduced-data elements to produce a nonlinear representation of the speech that includes as elements thereof the quantities proportional to the products and powers;   C. first modeling means for modeling the nonlinear representation with a group of modeling elements which include nonlinear representation characteristic of one or more speech-elements of interest in known speech, the modeling producing a reduced-data nonlinear representation that includes a plurality of reduced-data nonlinear representative data elements;   D. a second means for calculating quantities proportional to products of certain of the reduced-data nonlinear representation data elements and powers of certain of the reduced nonlinear representation data elements to produce a further nonlinear representation of the speech that includes as elements thereof the quantities proportional to the products and powers;   E. second modeling means for modeling the further nonlinear representation with a group of modeling elements which include nonlinear representations characteristic of the speech-elements of interest in known speech, the modeling producing speech-element data relating to the most likely speech elements present in the speech signal; and   F. means for consolidating the rearranging the speech-element data to produce a minimum speech-element representation of a word or phrase corresponding to the speech signal.   
     
     
       5. The speech-recognition device of claim 4, wherein the second nonlinear calculating means includes a nonlinear receptive processor which forms products of selected data associated with the energy of the speech signal in various frequencies and selectively combines the products to produce data which relate, respectively, to changes in the signal energy signal over various frequencies and over predetermined time periods. 
     
     
       6. A speech-recognition device for identifying a speech element of interest in a speech signal, said device comprising: A. means for generating a first vector each of whose components represents a component of said speech element;   B. means for comparing said first vector with a first set of model vectors representing known speech elements, for each comparison deriving a value representing the degree of correlation with one of said model vectors, and generating a second vector each of whose components is one of said values;   C. means for selectively calculating nonlinear combinations of components of said second vector, and generating a third vector having the resulting nonlinear combinations as components; and   D. means for comparing said third vector with a second set of model vectors representing known speech elements, the comparing means producing data which identifies the most likely speech-element present in the speech signal.   
     
     
       7. The speech-recognition device of claim 6, wherein the second nonlinear calculating means includes a nonlinear receptive processor which forms products of selected data associated with the energy of the speech signal in various frequencies and selectively combines the products to produce data which relate, respectively, to changes in the signal energy signal over various frequencies and over predetermined time periods. 
     
     
       8. The speech recognition device of claim 6, wherein said second set of model vectors corresponds to a predetermined set of phoneme isotypes. 
     
     
       9. The speech recognition device of claim 6, wherein said first set of model vectors corresponds to a predetermined set of phonemes. 
     
     
       10. A method of identifying a speech element of interest in a speech signal, said method comprising the steps of: A. generating a first vector each of whose components represents a component of said speech element;   B. comparing said first vector with a first set of model vectors representing known speech elements and for each comparison deriving a value representing the degree of correlation with one of said model vectors, thereby generating a second vector each of whose components is one of said values;   C. selectively calculating nonlinear combinations of components o said second vector, the resulting nonlinear combinations being the components of a third vector;   D. comparing said third vector with a second set of model vectors representing known speech elements; and   E. identifying as present in the speech signal the known speech element associated with the vector from said second set of model vectors to which said third vector most closely correlates.   
     
     
       11. A method of identifying a plurality of speech elements of interest in a speech signal, said method comprising the steps of: A. producing a first reduced-data representation of the speech signal segment that includes a plurality of reduced-data elements;   B. comparing the reduced-data representation with a group of modeling elements which include nonlinear representations which are characteristic of one or more speech elements of interest in known speech, and producing a second reduced-data representation with elements which correspond to the degree of correlation between the elements of the first reduced-data representation and the various modeling elements;   C. calculating quantities proportional to products of certain of he second reduced-data representation data elements and powers of certain of the second reduced-data representation data elements to produce a nonlinear representation of the speech that includes as elements thereof quantities proportional to the products and powers;   D. comparing the nonlinear representation with a group of modeling elements which include nonlinear representations which are characteristic of the speech elements of interest in known speech, said comparison producing a set of values which identify speech elements which correspond to the speech signal.